• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

GU-46 push-pull & single-ended amplifiers

Ex-Moderator
Joined 2011
@Thomas - If you do not need such high power, then the output tube selection seems somewhat odd... Using Douglas' suggested 10kp-p, with 200mA bias, the output power is 108W, while with 5kp-p and a hotter bias of 400mA, the output power is 181W, both are before OPT loss, which probably will be around 5~10%. These are the lowest output you can get, since both are running in Class A only. The load lines are as follow:

GU-46 Load Line.jpg

Pa_max is 500W per the datasheet:

Screenshot - 12_11_2017 , 12_55_53 PM.png
 
Hi Jazbo and Bandersnatch

Thanks for your support. So then es should probably go towards the 10k pp Version (due to lower Ampere). Or somewhere inbetween 10 and 5k.

For the tube specs/supply this would mean approx. -30V (how Mama A?) on g2, at approx. 1kv /800mA anode voltage?

What would this mean for the Power tranny, when we look at the combined proposal of Banat? Maybe it would make sence to separate the heater taps...

Thanks for your help up to now.
Thomas
 
A misunderstanding. Lowering the g2 voltage will lower the g1=0V line to the space occupied by the g1=30V plate current line. It is not g2 being -30V, it is adjusting g2 volts downward so that g1=0V has the same plate current as the g1=-30 does at 500V on g2.

5k a-a, with a 400 mA idle point will be fine, and at 80% of maximum, should keep the anode hot enough to do its duty as getter.
cheers,
Douglas
 
Ex-Moderator
Joined 2011
If you already have the GU-46, it would help a lot with the design process if you can trace its triode characteristic at much lower Va, Ia and Vg1 than what are shown on the Russian datasheet. With the triode characteristic, I can build a Spice model for you to simulate the output stage more accurately.
 
If that GU46 plate is Tantalum based than to not forget that filament burn around 125W which also steady heat the same plate on some percent .
That is allmost 250W for PP :eek:, for an amp of 160W out :stop:
If you go AB it looks more usefull.
Made Ia-Va cuves from the little info i found.So only an aproximation.
Mona
 

Attachments

  • Raa46.jpg
    Raa46.jpg
    49.7 KB · Views: 378
  • 46.jpg
    46.jpg
    33.4 KB · Views: 400
Hi Jazbo
I have the tubes here with me, but I do not have the material to test them in any mode yet.
I hoped to get the Info needed before ordering custom wound trannies...
I am afraid, it will most probably stay theory therefore.

It seems that my crazy and unfinished idea gives a lot of good people headaches... Sorry about that.
Thomas
 
Last edited:
If you go AB it looks more usefull.
Mona

Agree !

And I still think that AB1 class at Uba=1-1,2KV , Ug2=0,5-0,6KV,-Ug1=50-70V and Raa around 5K will make some optimum compromise efficiency wise , all that driven with slightly modified Mullard 5-20 front end .

btw , I guess that GU46 is more rugged power pentode than GU81M tube , need to found few pieces to try those tubes in my AM/RF transmitter ,class C, f=1,6-1,8Mhz , Ua=2,5-3,5KV , Ug2=500V ,g1=RF driven , g3= DC negative biased and AF modulated .

Regards
 
@Thomas - If you do not need such high power, then the output tube selection seems somewhat odd... Using Douglas' suggested 10kp-p, with 200mA bias, the output power is 108W, while with 5kp-p and a hotter bias of 400mA, the output power is 181W, both are before OPT loss, which probably will be around 5~10%. These are the lowest output you can get, since both are running in Class A only. The load lines are as follow:

View attachment 650626

Pa_max is 500W per the datasheet:

View attachment 650627

Jazbo, I think there is a wee bit more than 181 W before losses available from a PP pair at 1kV, 400 mA idle and 5k a-a load. By V^2 over R...the V across that 5kOhm primary is north of 1800V...right around 350W RMS output ahead of the load losses...:)
cheers,
Douglas
 
Agree !

And I still think that AB1 class at Uba=1-1,2KV , Ug2=0,5-0,6KV,-Ug1=50-70V and Raa around 5K will make some optimum compromise efficiency wise , all that driven with slightly modified Mullard 5-20 front end .

btw , I guess that GU46 is more rugged power pentode than GU81M tube , need to found few pieces to try those tubes in my AM/RF transmitter ,class C, f=1,6-1,8Mhz , Ua=2,5-3,5KV , Ug2=500V ,g1=RF driven , g3= DC negative biased and AF modulated .

Regards

If you get those worthless GU81m for an audio amp, you will wind up just like me: six pack of tubes, and a quad of sockets, and the discovery that they are just too much. Better to run || 813's...LOL
cheers,
Douglas
 
Ex-Moderator
Joined 2011
Jazbo, I think there is a wee bit more than 181 W before losses available from a PP pair at 1kV, 400 mA idle and 5k a-a load. By V^2 over R...the V across that 5kOhm primary is north of 1800V...right around 350W RMS output ahead of the load losses...:)
cheers,
Douglas
You are right, I used the wrong formula for the above calaculations, so please ignore them! Looking up RDH4, the correct formula is:

Po = 0.5 x Ia_max x Ebb, so for the 5kp-p load, Po is about 280W!
 
Hi Kay Pirinha
Thanks for your reply. I think I will go along with Banats proposal for a PSU as soon as the rest of the schematic becomes clearer.
Up to now, it's all theory, no parts, except the GU tubes I have here, have been ordered.
I am still completely open to proposals.
Unfortunately, I am a bit limited concerning measuring tools, I have no scope, which might make trouble shooting a bit difficult in complicated cirquits.
This is why I tried to stick to a phase splitter I know well and to a relatively simple driver. But If a Mullard 5-20 cirquit with EL34 tubes will do a good driver job for the GUs, why not.
Thomas
 
You are right, I used the wrong formula for the above calaculations, so please ignore them! Looking up RDH4, the correct formula is:

Po = 0.5 x Ia_max x Ebb, so for the 5kp-p load, Po is about 280W!

I have to ask what are your numbers? I use your formula and get 360W...which is pretty close to the way I calculated it and got 350W...:)

I make Ia_max at .72A btw.
cheers,
Douglas